Product
Error Explainer
Turn stack traces, failed commands, and visible logs into a probable cause and clear next step.
What it does
Error Explainer reads visible errors, warnings, or log output, then returns the likely cause, the most relevant context on screen, and the next debugging move worth trying.
Input
- the captured primary-screen screenshot with the visible error or logs
- the spoken request, usually something like "what is wrong here?"
- optional follow-up context if the first answer needs more detail
Output
- a concise spoken explanation
- a markdown card with cause, evidence, and suggested next steps
- copyable debugging instructions or a likely fix
- an optional clarifying question if the screenshot lacks enough context
Why it is useful
The mode removes the tedious part of debugging triage: reading a noisy stack trace, summarizing it, and deciding what to check first. The user stays in the existing editor or terminal instead of switching into search mode.
Current implementation intentionally does not do
- run code, tests, or commands automatically
- inspect files outside the visible capture
- claim certainty when the screenshot is ambiguous
- behave like a general autonomous agent